Two uses of chance, two different questions
A study might use the word “random” without making clear what was randomized. That detail matters. Random sampling is a way to select people or other units from a defined population. It addresses the question, “Who gets into the study?” Random assignment is a way to allocate participants already in a study to different conditions. It addresses the question, “What does each participant experience?”
Imagine a researcher wants to compare two versions of a reminder message. They could invite volunteers to take part, then use a random process to decide who receives each version. That is random assignment, but not necessarily random sampling: the volunteers were not randomly selected from everyone the researcher might hope to understand. Alternatively, the researcher could randomly select people to survey about reminders without asking them to receive either version. That would be random sampling without random assignment.
Neither method is automatically present just because a study has a large number of participants or uses software. The useful habit is to find the precise decision that chance controlled.
What random sampling helps you understand
With random sampling, each eligible member of a specified population has a known chance of selection under the sampling plan. This helps researchers use information from a subset to describe the larger group it came from. The population must be named carefully. A random sample drawn from a customer list, for example, does not automatically represent everyone who might buy the product.
Selection is only part of the story. If many selected people do not respond, or if the list leaves out an important group, the final set of responses may differ from the intended population. Wording, measurement and timing can also affect what a survey tells us. Random sampling supports population-level estimates under appropriate conditions; it does not, by itself, show that one thing caused another.
When reading a study, ask: Who was eligible to be selected? What list or process was used to reach them? Who actually participated? The answers are more informative than a bare claim that the sample was random.
What random assignment helps you compare
Random assignment places participants into conditions by chance rather than letting them or the researcher choose. In the reminder example, participants might be assigned to receive version A or version B. The goal is to make the groups comparable before they receive the messages, so differences observed afterward are less likely to reflect pre-existing differences between the groups.
This is why random assignment can strengthen a cause-and-effect conclusion. If the groups were treated similarly apart from the message, a difference in their outcomes provides evidence about the effect of the assigned message. It is not a guarantee of a clear answer. Chance can still produce uneven groups, especially in a small study, and missing outcomes, poor measurement or other differences in how the groups were treated can complicate interpretation.
Assignment also needs a clearly defined comparison. A study comparing two reminders can speak to the difference between those reminders in that setting. It cannot, without further evidence, establish how either reminder compares with every possible alternative.
A study can use both, one or neither
Suppose the researcher randomly samples people from a membership list and then randomly assigns those who take part to the two reminder versions. Sampling helps connect the participants to the list’s population; assignment helps compare the effects of the versions among participants. Together, the methods can support a stronger argument about both the comparison and its relevance to that population, provided participation and study procedures do not undermine either step.
Now suppose the participants are volunteers but are randomly assigned. The comparison may still offer useful evidence about the effect for people like those volunteers. Applying it to a broader population requires more care. If instead people are randomly sampled but choose which reminder they receive, a difference between the groups could reflect their preferences or circumstances rather than the reminder itself.
Many useful studies use neither method. An interview study, for example, may deliberately recruit people with varied experiences to explore how they understand a topic. Its value should be judged against its purpose, not by whether it resembles an experiment or a population survey.
A simple way to read the methods
Draw two arrows in your mind: one from the target population to the participants, and another from the participants to the study conditions. For the first arrow, ask how people were selected and who was missed. For the second, ask whether chance determined the conditions and whether the groups were otherwise treated comparably.
Then match the conclusion to the method. A claim about how common an opinion is in a population calls for attention to sampling and participation. A claim that a message changed behavior calls for attention to assignment and the comparison conditions. If a report makes both claims, examine both arrows. The distinction turns “Was it random?” into a more useful question: “What was random, and what does that allow us to conclude?”
